Comparative Study of Feature-based Algorithms and Classifiers in Face Recognition for Automated Attendance System

Attendance plays an indispensable role in any educational system. In this research work, an automated attendance recording systems has been proposed to update the attendance records of the students by recognizing their physical presence in the classroom with the help of face detection and recognition techniques. Face recognition mimics the operations of object recognition in computer vision and image processing domain. PCA and LDA are the two widely used models for feature extraction in face recognition algorithms to extract the low dimensional and more discriminating features from face. The purpose of this paper is to present an autonomous, relative study of three state-of-the-art appearance-based feature extraction methods (PCA, LDA and Hybrid approach) in completely even conditions regarding processing and algorithms execution. The experiments were performed on face databases known as SDB (Student Data Base), where the images were collected by using Truevision HP laptop camera with different illumination conditions at different angles with different facial expressions and poses etc. The hybrid approach improves the face recognition rate when compared with PCA and LDA using SVM as a classifier.